Skip to content
Agent Observability Platform
Open Source

AgentLens

See why your AI agents make the decisions they make. Complete observability for multi-agent systems.

pip install vectry-agentlens

Everything you need to understand your agents

From decision trees to cost intelligence, get complete visibility into how your AI systems operate

Decision Trees

Visualize the complete reasoning behind every agent choice. See the branching logic, alternatives considered, and the path chosen.

Context Awareness

Monitor context window utilization in real-time. Track what's being fed into your agents and what's being left behind.

Cost Intelligence

Track spending per decision, per agent, per trace. Get granular insights into where every dollar goes in your AI operations.

Quick Setup

Up and running in 3 steps

Go from zero to full agent observability in under five minutes

1

Install

One command to add AgentLens to your project.

pip install vectry-agentlens
2

Instrument

Add the @trace decorator or use wrap_openai().

@trace / wrap_openai()
3

Observe

See every decision in the real-time dashboard.

agentlens.vectry.tech
Python SDK

Instrument your agents in minutes

Our Python SDK integrates seamlessly with LangChain, OpenAI, and custom agents. Add observability without changing your architecture.

  • Drop-in decorators for automatic tracing
  • No code changes required for LangChain
  • Async-first design with minimal overhead
  • Type-safe with full IDE support
decorator_pattern.py
@trace
from agentlens import init, trace

# Initialize AgentLens
init(
    api_key="your-api-key",
    endpoint="https://agentlens.vectry.tech"
)

# Trace your agent function
@trace(name="research-agent")
async def research(topic: str):
    # Your agent logic here
    return f"Researching: {topic}"
openai_wrapper.py
wrap_openai
from agentlens.integrations.openai import wrap_openai
import openai

# Wrap the OpenAI client
client = openai.OpenAI()
client = wrap_openai(client)

# Use as normal - traces automatically
response = client.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": "Hello!"}]
)
Integrations

Works with your stack

First-class support for the most popular AI frameworks. Drop in and start tracing.

OpenAI

GPT-4, GPT-3.5, o1

wrap_openai()

LangChain

Chains, Agents, Tools

Auto-instrumented

Custom Agents

Any Python agent

@trace decorator
Pricing

Simple, transparent pricing

No hidden fees. Start free, scale as you grow. Every plan includes the full dashboard experience.

Free

For experimentation

$0/month
  • 20 sessions per day
  • Full dashboard access
  • 1 API key
  • Community support
Get Started Free
Most Popular

Starter

For small teams

$5/month
  • 1,000 sessions per month
  • Full dashboard access
  • Unlimited API keys
  • Email support
Start Starter Plan

Pro

For scaling teams

$20/month
  • 100,000 sessions per month
  • Full dashboard access
  • Unlimited API keys
  • Priority support
Start Pro Plan